I'm interested in reproducing the crash, do I only need the gfx.wad and do you have a d/l link?

Thanks for your generosity :) http://www.phait-accompli.com/q/s4/s4pak.zip 300 something KB.

One other problem (I hate posting all my problems in the midst of other's, but bear with me) -- I recently full VIS'd a map, which has no critical errors. It runs fine, I can quicksave and normal save. But when I quickload or normal load my point in the map - it gets stuck at "Loading" and I have to go to the console. In the console it says:

Loading game from e:\games\quake/id1/quick.sav...
couldn't spawn server maps/systkills:__0/__4____.bsp
couldn't load map

What sticks out here is the slashes - both \ and / is used... this problem occurs in FitzQuake 075, 080, and even WinQuake - and my friend experienced the same problem. I've never seen anything like it! You would think there is some map name problem, but the BSP is s4m1.bsp - and the title is something like:

Storage Facility E9

System 4
By Phait 

be the old "newlines in title" issue; take a look at my ToolTips document, Engine section. My engines can load such a savegame but the general solution is to remove the newlines from the title.

Thanks for the pak, I'll check it out. 

I'll look into the newlines stuff, that was one of my suspicions but didn't make sense as I've played maps w/ such titles and had no save/load problems. Thanks. 

Thanks for the help, sorry I missed it before. I guess I deserve a spanking or something. 

Reading the Q1Tooltips, I'm confused as your advice regarding the BSPInfo. I've extracted the .ENT file just fine, but: and then rebuild using qbsp with the "-onlyents" option.

How does the BSP update with the edited .ENT file? I tried qbsp -onlyents map.map map.bsp - but that didn't update it. 

I searched up the error, and found this in the Quake source:

/*
=============
Draw_TransPic
=============
*/
void Draw_TransPic (int x, int y, qpic_t *pic)
{
if (x < 0 || (unsigned)(x + pic->width) > vid.width || y < 0 ||
(unsigned)(y + pic->height) > vid.height)
{
Sys_Error ("Draw_TransPic: bad coordinates");
}

Draw_Pic (x, y, pic);
}


/*

That may help. Also, while testing this error again I clicked somewhere outside of the dialog box, I believe and got this error: Double Quake error: R_RenderView: called without enough stack - whic you can find in http://www.meowfishies.com/code/quakeforge-patches/files/r_main.c 

I'd love to check it out, but I'll have to wait until monday when I'm back at my apartment.

But: 1. What is the error message when it crashes? 2. Does it work in GLQuake? 

GL Quake just freezes after "CD Audio Initialized", I can't access any menus or type anything.

Here is the Fitzquake error:
http://www.phait-accompli.com/q/fqcrash.jpg

I am certain it is the same problem I'm experiencing in WinQuake, except FQ doesn't display the 2 errors I get in WinQuake - just a cryptic crash. 

I just re-read your thing that answers #1.

BTW, I think your problem with winquake is that the graphic is so big tht it goes off the screen. Winquake can't handle that situation, if i recall correctly. 

well i'll try to reproduce it next week. It might be an access violation that windows is helpfully generalizing into "an error." 

Well, the thing is I am not changing the original graphic's dimensions, only the look. 

I thought it would work so, but by making one dxf frame in QMLE and saving it to mdl gives a rather strange map that won't load in Quark.
I tried the AC3D studio, and this made a Quake map.

Thing is that all the polyhedrons-strings were after each other, while in a quake map there are just 6 strings to define a polyhedron.
So I diveded them in 6 but still Quark crashes.

Thanks HeadThump. That could be a solution. Fact is that I only can work with Quake1.
Q2 and Q3 I never compiled for, so if I would try that I should make myself convienent with those compilers. And I once made some things with blender, but I can't say I can work with it. 

The engine crash bug loading the System4 gfx.wad is caused by a texture anum_0 with invalid dimensions. You can see this also by loading the wad in TexMex; it will refuse to load that texture. Many GL-engines will crash in this situation, but DarkPlaces (Oct 2004) reports correctly.

I've now added checks and messages with texture name to my engines.

As for the onlyents issue; just run:
qbsp -onlyents mapname.ent

If you don't specify extension, qbsp will look for the default extension .map and in this case, it doesn't contain the new changes.

OTOH since I assume that you have the map source, I'd recommend change that instead and rebuild, instead of using the BspInfo method. Use that only if you don't have the source. 

Thanks very much for your help :) 

Are you sure? Because I loaded the wad in Tex Mex, and anum_0 shows just fine. It's CONCHARS that never loads (whether in my wad or id's original wad). 

I replaced my modified anums with the original and it now loads. I won't doubt you ArguiRe ;) 

I misinterpreted TexMex' behaviour a bit. The wad is still invalid (including the anum_0 tex report from the engine) but as you say, it's the conchars tex that's not loaded.

After checking some more, I think I have an idea what the problem is. I read somewhere that the original gfx.wad is in fact invalid because they made some hackish addition to the wad and patched the header. This is why TexMex has problems loading the wad. You can check loading e.g. the Hipnotic gfx.wad instead and it will load properly.

However, there's also another problem. You seem to have created these new objects as mip textures (with mipmaps) and that's wrong. They should be StatusBar Pictures instead and definitely without mipmaps.

I've tried to figure out how to convert an existing mip texture to a StatusBar object (i.e. remove the mipmaps) but haven't had any success yet.

Maybe someone else knows how to do this? 

That my graphics have become MIPs is involuntary. What I did was export all the graphics to .PCX, opened in Photoshop and edited there. Then, whichever ones needed to be .LMP were converted. The ones inside GFX.wad were merely imported .PCX files. How they got turned into MIPs, dunno. Must be automatic. 

Dunno where one would get that nowadays tho... 

Use the evil folder deleting tool (or Adquedit) for importing into your gfx.wad. It should always save it correctly, it accounts for the wierdness of the conchars file and all that automatically. 

I have decided to use the 'sinking' monster route to get rid of my corpses.

I have created a simple SUB_Corpse_sink that lowers the origin_z value a unit at a time and lowers my corpse until it is out of sight. I then use remove(self). It works so far.

However, if I quickly no_clip to the floor below, I can see the corpse appear through the ceiling before suddenly disappearing altogether.

I am thinking that if a monster is on a walkway above the player and is then killed, the player could see all this happening. Is there a way to get the player to sink into the 'ground' by (effectively) removing it in layers. I have in mind a pack of cards that I keep removing the bottom card until the pack has gone.

I have seen this sinking effect in Lost Chapters although I do not know how it is achieved there. 

The corpse removal in Chapters works exactly as you appear to have implemented it yourself. If you check out any monster that dies on a walkway in Chapters ( the ogre above the GL in chapter_secret for example ) you'll see his corpse sink through the wooden boards and turn black before disappearing.

Preventing the model showing through is basically impossible in Quake as far as I can tell.
Trying to skim the bottom off the model like a deck of cards would require that the polys be rendered individually and relative to world polys, which Quake simply doesn't do. I think that would need an engine mod. DP could maybe do it.

re: chapters progs. No word from necros, guess I'll have to email him again. Unless he responds here. 

OK, thanks. I'll carry on the way I'm going. 

What you could do is make an extra animation sequence where the model itself sinks down, but that any vertex that is more than 2 quake units below the floor gets flattened. That way nothing would show through a floor more than 2 quake units thick. Of course, doing that kind of animation wouldn't be smooth on an engine without interpolation, so perhaps what you want is to move the model down as you animate it, flatten the vertices at -2, undo the motion downwards, then code it to move at the same rate you originally animated it at. Urrgh. But it should would.

Except...this kind of calculation is assuming the corpse is on a single flat surface. None of this would be valid on a slope. There would be work arounds to minimise that problem, like doing a traceline down, and using the trace_plane_normal to calcluate how far to put it down before playing the animation. And nothing you can do will help if the corpse hangs off the edge of a ledge. So yeah, it's never gonna be great, but it could work better 

nt